Skip to content

Latest commit

 

History

History
61 lines (55 loc) · 2.36 KB

README.md

File metadata and controls

61 lines (55 loc) · 2.36 KB

digital-defender

This is the repo to collaborate on Thorn-Cloudera's project at OSD- GHC16. Please create a folder with your team name, add your work there and submit a pull request.

Here is the project description page: http://systers.org/wiki/communities/doku.php?id=wiki:ghc:ghc16:osd:cloudera

Data Sets:

Sanitized Chat Data over a few months for 2 chat rooms

{
‘author': 'User1',
'content': '@User2 MESSAGE LINK1 ',
'id': u’ef66e354-aa47-42fc-92f3-dcbed2802510',
'scrape_datetime': u'2016-09-06 02:10:31.152588’,
'site': ‘SITE1'
}

Guiding Questions

  1. UI/UX
  • Landing page show high level overview: chat room stats, top user table, includes search
  • Views: user profile, content, network analysis
  1. Data Insights
  2. Messages over time per chat room
  3. User Stats: * distributions by: # messages, # unique links posted * Top Users: * post patterns, num of original links posted
  4. Content Stats: * Most popular content, distribution of content by popularity * How often is new content posted?
  5. Network * Who talks to who? graph analysis to rank users? do sub communities exist?

Tools (Need to finalize once we decide on our data hosting and analysis strategy)

Javascript, React, Python (numpy, scipy, pandas, networkx), Jupyter notebooks, git

Open questions:

  1. Which bootstrapping model do we want to choose for data hosting and analysis?
  2. Option 1: Provide the dataset in a downloaded location And participants can download, use their favourite tools for exploration and generating stats. Upload the code and insights to github.
  3. Option 2: Hosted on Cloudera infrastructure * Create an EDH cluster - Sravya * Create Thorn dataset as a Hive table? - Sravya * Make sure Spark is setup and we can run spark jobs on this dataset * User account setup: Setup cluster accounts for all users. * Analysis:
    • Spark shell
    • Hue console
    • Sense
    • Jupyter notebook * Preparation before hackathon is strongly encouraged:
    • Get familiar with one of the analysis tools mentioned above.
  4. Option 3: Docker?
  5. We do not have mentor expertise in react and javascript
  • Can we get more mentors to fill this expertise gap?
  • If not - shall we remove this project?

Data Hosting? (Sravya - I am not sure what this is?)

LINK1, FINGERPRINT 1 LINK2, FINGERPRINT 1 LINK3, FINGERPRINT 2